基于CMAC的触觉传感信号自适应逆滤波器设计

CMAC是模仿人类小脑工作机理的一种神经网络,它具有自适应补偿能力。本文首次在机器人PVDF触觉传感器领域使用了基于CMAC神经网络的逆滤波系统,目的是消除在通过触觉传感器的应力信息反求被触摸物体的轮廓过程中系统模型产生的结构误差。使用ARMA模型对触觉传感器模型进行了较好的逼近,保证了整个系统能有效工作。设计了基于频域特性的物体轮廓高程估计逆滤波器系统,此逆滤波器具有较快的速度和很强的通用性。

Design Tactile Sensor Signal Self-Adapt Inverse Filter Based CMAC

CMAC is a kind of neural network which simulates human cerebellar mechanism,which has the self-adapt com-pensation function.This paper may be the first to apply a CMAC network inverse filter system in robot PVDF tactile sen-sor field.The system effectively eliminated the structure error which is produced in the process of converting tactile sensor stress signal to object outline.An ARMA tactile sensor approx-imate model was also applied to make the system efficient.In-verse filter system is designed in frequency space.The filter has high speed and is universal.